Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: save std. error, confidence interval, svymean


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: save std. error, confidence interval, svymean
Date   Mon, 20 Sep 2004 14:51:29 -0400

--- In statalist, Andreas Stiehler wrote:
> do you know a way to save Std.Error and Conf.interval of a "svymean"  
> operation (see below) as matrices or scalars?

-svymean- is an estimation command, so it leaves a bunch of stuff in
the -ereturn- values. This is my weird example with auto data (I was
asking for a "standard" toy data set with complex enough sampling
design, but don't remember if any such data set in public use actually
arose then):

. sysuse auto
(1978 Automobile Data)

. svyset [pw=turn] , psu(rep)
pweight is turn
psu is rep78

. svymean pri

Survey mean estimation

pweight:  turn                                    Number of obs    =        69
Strata:   <one>                                   Number of strata =         1
PSU:      rep78                                   Number of PSUs   =         5
                                                  Population size  =      2746

------------------------------------------------------------------------------
    Mean |   Estimate    Std. Err.   [95% Conf. Interval]        Deff
---------+--------------------------------------------------------------------
   price |   6251.806    204.7158    5683.423    6820.188    .3211447
------------------------------------------------------------------------------

. eret li

scalars:
              e(df_r) =  4
                 e(N) =  69
          e(N_strata) =  1
             e(N_psu) =  5
             e(N_pop) =  2746

macros:
               e(cmd) : "svymean"
           e(predict) : "svy_x_p"
           e(varlist) : "price"
          e(complete) : "complete"
            e(depvar) : "Mean"
               e(psu) : "rep78"
              e(wexp) : "= turn"
             e(wtype) : "pweight"

matrices:
                 e(b) :  1 x 1
                 e(V) :  1 x 1
              e(V_db) :  1 x 1
               e(est) :  1 x 1
             e(error) :  1 x 1
                e(_N) :  1 x 1
           e(_N_subp) :  1 x 1
             e(V_msp) :  1 x 1
             e(V_srs) :  1 x 1
              e(meft) :  1 x 1
              e(deft) :  1 x 1
              e(deff) :  1 x 1

functions:
            e(sample)   

. di _b[price]
6251.8055

. di _se[price]
204.71576

. mat li e(b)

symmetric e(b)[1,1]
        price
y1  6251.8055

. mat li e(V)

symmetric e(V)[1,1]
           price
price  41908.543

. mat li e(deff)

symmetric e(deff)[1,1]
        price
r1  .32114473

So you can save them for later use as

scalar mean = _b[price]
scalar semean = _se[price]

etc. More advanced methods of dealing with estimation results are due
to Roger Newson and his idea of estimation sets. -findit parmest- to
find more.

-- 
Stas Kolenikov
http://stas.kolenikov.name
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index